Timeline for Mapping 3D object coordinates to 2D screen coordinates with perspective projection
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 23, 2013 at 15:49 | comment | added | Paul Riding | I think the matrix multiplications are in the correct order. I'm not entirely sure but I think it has something to do with the last division of the x and y components by the z. Here's my most recent Camera3D class where I organise the matrices and render the points. Can you spot any errors that would cause the points to render in the opposite corner after passing the Z value of the camera (behind): pastebin.com/NGGHpqGv | |
| Feb 21, 2013 at 11:58 | comment | added | ScrambledRK | hmm, maybe you can resolve this transposing a few matrices or changing the multiplication order. its a problem I ran into a lot, as its difficult see if the code is column or row based. | |
| Feb 21, 2013 at 1:45 | comment | added | Paul Riding | This is great but the problem I'm having is when the Z component is negative and supposedly out of the frame and behind the camera, it appears to just come back into the screen from the opposite corner of the screen. e.g. A point at coordinate 1:1:0 should be in the top right corner of the screen. If I increase its Z component it moves forward as expected, but if I decrease it, rather than moving behind the scene and not being rendered, it begins to move in from the bottom left hand corner towards the centre of the screen, in this example. | |
| Feb 20, 2013 at 22:34 | vote | accept | Paul Riding | ||
| Feb 20, 2013 at 22:34 | vote | accept | Paul Riding | ||
| Feb 20, 2013 at 22:34 | |||||
| Feb 20, 2013 at 8:31 | history | edited | ScrambledRK | CC BY-SA 3.0 |
added explainaition for normaldevicespace and clipspace
|
| Feb 20, 2013 at 8:21 | history | answered | ScrambledRK | CC BY-SA 3.0 |